1. Field of the Invention
This invention relates to apparatus and methods for monitoring I/O performance, and more particularly to apparatus and methods for determining the cause of degraded I/O performance between a host system and a storage controller.
2. Background of the Invention
Storage networks, such as storage area networks (SANs), are frequently used to interconnect different types of data storage systems with different types of servers (also referred to herein as “host systems”). In many cases, the host systems and storage systems are tuned (i.e., configured) to optimize I/O performance between the host systems and storage systems. Modifying the configuration or settings of the host systems and/or storage systems has the potential to negatively affect I/O performance between the systems. Furthermore, degraded I/O performance between any particular host system and storage system can potentially affect the I/O performance between other host systems and storage systems connected thereto, thereby generating system-wide I/O performance problems.
In certain cases, a host system communicating with a storage system may include a performance measurement tool to monitor I/O performance between the host system and the storage system. For example, Resource Measurement Facility (RMF) is a performance measurement tool for z/OS-based host systems that monitors and records performance statistics between the host system and a connected storage system. Although tools such as RMF may be effective to measure and detect degraded I/O performance, such tools are limited in their ability to determine the cause or source of the degraded I/O performance. In certain cases, tracking down the cause or source of the problem may require many man-hours analyzing data from the affected host systems and/or storage systems. In worse case scenarios, determining the cause of the problem may involve recreating the problem with sniffers on both hardware and software in an attempt to determine the root cause.
In view of the foregoing, what are needed are apparatus and methods to more effectively determine the cause of degraded I/O performance between host systems and storage systems. Further needed are apparatus and methods to monitor configuration changes to host systems and storage systems the can potentially affect I/O performance. Yet further needed are apparatus and methods to notify host systems when I/O performance with a storage system has been degraded in a significant manner.